Robust Design for Systems or Processes

A key to recent success in the manufacturing industry has been the application of Robust Design or Taguchi Method. While the details of the process get deeply into statistical mathmatics, the concepts are quit simple.




This diagram demonstrates the key relationships between any process or system and its environment. The "signal factors" are the inputs into the system, this may include a product at a certain stage in a process and then there is the "response" which is the condition of the product after it has been through the process or system. For this "black box" there are also system inputs which includes "noise", things that can't be controlled, and "control Factors" things which can be controlled.

As system operators, we operate many systems where we attempt to control the process or system with our control factors. The lesson that we learn from Robust Design process is that these control factors often have a complex and difficult relationships. Many times we attempt to control a process or system response by varying certain control factors and ignoring others, without any understanding of how these processes relate to each other. Robust Design when applied allow you to determine the "Control Factors" which have the most leverage in the control relationship.

Incidentally, these "Control Factors" would make great Key Performance Indicators (KPI)'s for controlling your process or system.

Robust Design is not often discussed in the context of Asset Management, however, using the statistical strategies to evaluate failure data could be very beneficial in determining techniques that will allow significant extension of life.